To Pivot or Not to Pivot: How to Know What Needs to Change

A few years ago, a young company had built a relatively simple way for factories to see something they had rarely been able to see clearly before: how individual machines were consuming electricity.

The original goal was sustainability. Put inexpensive sensors on factory equipment, expose its energy use in real time, and help manufacturers cut waste and emissions.

The technology worked, but the company struggled to generate healthy sales. Demonstrating strong product-market fit around energy management alone was elusive.

Then they noticed something: Customers were using the same electrical signals for something different. A machine’s power usage reveals not just energy use, but also its productivity.

Factory managers cared about this information deeply, because downtime and lost capacity hit the bottom line directly.

Very little of the technology changed, but the problem now being solved was fundamentally different.

Guidewheel started with an energy-sustainability mission in Kenya, but customer behavior pointed to a much bigger opportunity: machine performance and productivity.

What changed was the benefit customers cared about more, which led them to a much better answer.

STEM founders in particular, face long horizons, high capital needs and fuzzy signals from the market. That kind of ambiguity makes pivoting especially hard for STEM founders, where feedback is slower and each wrong turn is expensive.

This week: how to tell when to keep going, when to change direction, and when to spend a little more to find out.

The usual advice on pivoting is pretty familiar: if the evidence says your current approach isn’t working, change direction; if it is, keep going. In The Lean Startup, for example, Eric Ries famously framed the choice as “pivot or persevere”, that is, make a substantive change to the strategy, or stay the course based on what you’re learning.

But there’s a problem if you’re a founder in STEM: the decision is rarely that clear.

Say you’re making progress but it’s too slow, or not giving you the exact evidence you need. That could be because of many reasons:

• The idea itself is wrong

• The idea is right, but it’s taking longer than you expected to prove out

• The technology works, but the execution, channel or customer choice is getting in the way.

• The idea is still right, but proving it may now cost more than it’s worth, or more than you have cash for.

When this happens, what do you do next?

The answer isn’t always to keep going or to pivot. Sometimes the better move is to pause before committing much more money and figure out exactly what you still need to know.

If you think about it, investors face exactly this problem all the time. It’s worse for them because they are one step removed from you, and don’t have your intimate understanding of the technology and the market. But they still have to make important capital decisions.

In a classic Journal of Finance study of 794 venture-backed companies, Paul Gompers found that VCs staged their capital and monitored companies more closely when uncertainty was higher. They didn’t wait until every question was answered, but they also didn’t commit all the money upfront. Each round gave them a chance to see what had been proved before deciding whether to put in more.

As a founder, you can use this same proven logic with your own runway. Before spending heavily on the current path, or jumping prematurely to a new one, what if you spent the smallest reasonable amount to answer the question that is holding up your decision?

That gives you a third option between pivoting and persevering: probing.

The Unlock

When progress is unclear, don’t start with “should we pivot?” Start with the pillar assumptions that must be true for this business to work: this technology can meet the required standard, the selected customer segment cares enough to act, you can reach them reliably, and the economics can work.

For the weakest assumption, identify what you need to prove next, and what that will cost in time and money. If that proof takes more time and money than you have, generate other options for that pillar, for example, a different application for the technology, a different customer segment, a changed sales process, or a cheaper raw material.

Choose the option that preserves the value of what you’ve already proved, and gives you a realistic next proof point. If you can’t find one, then you may need to rethink the product, market, distribution and profit from the ground up.

Say you’ve been building a medical device, but every time you test it in hospitals, it still doesn’t perform well enough, even after several attempts, and you’re starting to wonder whether to keep pushing or try something different.

Here’s how you might think through the problem:

1. What has to be true for my current path to work?

The technology can hit the required performance threshold; hospitals care enough about the problem; they can actually adopt it; and there is a viable economic model.

2. Which of those is actually in doubt?

You know hospitals clearly care and the economics look fine. But you’re not sure about technical performance. That tells you that you don’t yet have a “business isn’t working” problem. You have a specific technical uncertainty.

3. What would I need to see to justify continuing?

You know that the industry standard for similar devices is 90% accuracy. Your device accuracy is 85%. Now you have a decision threshold.

4. Can I get that answer within the runway I can afford?

If you can run the test in six weeks for $40K, you have your answer. But if proving it requires another 18 months and $2 million that you cannot realistically raise, the technology does not necessarily fail, but your current path may no longer be viable.

5. What exactly should change if the answer is no?

If your device can only demonstrate 85% accuracy, perhaps that is inadequate for surgery but perfectly valuable for veterinary applications. Don’t throw away the technology because the original application failed. Change the assumption that failed.

From The Capital Playbook

PROVE THE CUSTOMER WILL ACTUALLY BUY.

First confirm that you have the right customer and a problem important enough to change behavior and pay for. Look for action: a paid pilot, budget approval, signed agreement or repeated use.

If the need is real but purchases still stall, find where: wrong buyer, ineffective channel, procurement, workflow change, regulation, integration, implementation or pricing. Fix the specific barrier before changing the whole business.

CHECK IF THE TECHNOLOGY FITS THE APPLICATION.

Define the minimum real-world performance the application requires—accuracy, reliability, cost, durability, scale, safety or whatever matters most—and compare it with what you have demonstrated.

If the gap is bridgeable, price the next technical proof. If it is large or unrealistic, look for an application where the technology already meets the bar.

CHECK IF THE ECONOMICS WORK.

Test whether the price, gross margin, sales cost, implementation cost and sales cycle can support a viable business.

Customers may want the product and the technology may work, but if every sale takes 18 months and heavy customization, you may still have the wrong application or business model.

IF YOU PIVOT, FOLLOW THE PROOF YOU ALREADY HAVE.

Don’t begin with “what else could we build?” Ask where the technology already meets the required standard, which customers have shown greater urgency, and which technical results, IP, data, manufacturing work and customer learning would still apply.

The best adjacent market is not simply the largest. It is one where there is evidence of demand and much of what you have already proved still counts.

The One Thing

If progress is slowing, a raise is dragging, or runway is getting tight:

1. Look across the four load-bearing assumptions: customer, technology, market access and economics.

2. Pick the one with the weakest real evidence. e.g., no one has paid; only lab validation; no repeatable buying path; unit economics still do not work.

3. Pinpoint exactly what still needs to be proved. E.g., $25K willingness to pay; 95% field accuracy; a purchase completed within six months; 60% gross margin.

4. Specify what you actually know today. Who has paid? What standard has been met? What part of the buying process has been completed? Which number is hurting the economics most?

5. Find the biggest blocker. No budget owner; one unresolved technical limit; procurement stalls after pilot; manufacturing cost is too high.

6. Choose the cheapest test that could unblock progress: Ask the strongest prospect for a paid pilot; test the critical technical threshold in the field; get one live customer through procurement approval; change price or manufacturing cost and measure the margin impact.

Build wisely!